Obstructive sleep apnoea (OSA) is associated with a variety of important complications, namely cardiovascular, neurocognitive and metabolic disturbances. The prevalence of OSA is well studied in children and adults. However, adolescence - an interface between childhood and adulthood, and a period of developmental changes known to affect sleep is largely unexplored in relation to OSA. The only published prevalence study on adolescents is limited by its small sample size, not a true representation of the general population and primarily focused on Caucasians. In this proposal, the investigators aim to determine the prevalence of OSA, and associated clinical features in a population-based sample of adolescents aged between 12 and 16 years.

The sample selection will be based on a stratified (by districts) and clustered (subjects within randomly selected schools) randomised sampling frame. Each participant will fill in a sleep habit questionnaire, undergo anthropometric measurements, physical examination and complete home polysomnographic recordings. Participants will undergo Conners' Continuous Performance Test and have blood samples taken to phenotype their cardiovascular and metabolic risk. The primary outcome is prevalence of OSA, assessed by the obstructive apnoea hypopnoea index. Secondary outcomes include use of logistic regression models to assess association between different severities of OSA and various demographic, clinical and laboratory variables.

The obtained result will provide the much-needed OSA prevalence in adolescents which is essential for estimating the true burden of disease within this population. This information is also vital when considering population-based health policies and interventional strategies. Globally, the findings from currently evidence-sparse region of the world allow future international comparison of disease burden. Our study will also form a platform from which repeated measurements can be made to assess time trends and to answer the important question of whether adulthood OSA takes its origin from adolescence.

Study Design

It is a cross-sectional, population-based study. Subjects who are found to meet all the inclusion and none of the exclusion criteria will be recruited. Participants will complete a set of sleep-related questionnaires, undergo anthropometric measurements, home polysomnographic (PSG) monitoring, attention assessment and phenotyping of their metabolic and cardiovascular risk.

Collection of Data The subjects will attend our unit for assessment and explanation of home PSG monitoring. After a 15-minute rest, systolic (SBP) and diastolic (DBP) blood pressure will be taken using the Datascope Accutorr Plus. Two measurements are taken at 1-minute intervals, with a third reading taken 5 minutes later if the difference between the first two >4 mmHg. The average of two repeatable readings (difference ≤4 mmHg) will be used. Anthropometric parameters including standing height (in metres) without shoes and body weight (in kilograms) will be recorded using Harpenden Stadiometer (Holtain, UK) and an electronic weighing scale (Tanita BF-522, Japan) respectively according to standard recommendations. Waist and hip circumference (in centimetres) will also be obtained. Neck circumference will be measured at the level of the most prominent portion of the thyroid cartilage with the head held erect and the eyes facing forward. Z scores of body mass index and waist circumference will be calculated using local references.

If the participating school is agreeable and logistically feasible, our research team can carry out the above at school. However, in our experience schools close to our unit prefer to send their students to us to undergo the various assessments.

All assessments will be carried out during the week to avoid weekday-weekend difference in adolescents' sleep pattern. The students will also complete a chronotype preference questionnaire, result from which can provide guidance regarding effects of sleep phase. The reduced version of Horne and Östberg Morningness-Eveningness Questionnaire (rMEQ) will be used to measure chronotype preference. It consists of five items designed to assess one's preferred bedtime and rise time, the degree of tiredness in the morning, peak time for optimal performance, and self-perceived circadian preference. The rMEQ has been validated in ethnic Chinese individuals, with good psychometric properties [36], and has been used in previous research conducted in adolescents.[37, 38] Participants scored higher than 17 and lower than 12 will be classified as morning-type and evening-type, respectively. Participants with total rMEQ scores between 12 and 17 will be classified as intermediate-type.

Sleep study

Level III sleep study will be conducted at home for every subject. A model of NOX-T3 sleep monitor will be used to record the following parameters: respiratory movements of the chest and abdomen by inductance plethysmography, oxygen saturation and derived heart rate by oximetry, respiratory airflow by pressure transducer, snoring, body movement and body position. An adequate overnight sleep study is defined as one with total sleep time >6 hours.

A random sample of subjects will also be invited to undergo overnight full polysomnography (PSG) in our sleep laboratory to validate the level III study (please see below for details). In our laboratory, a model SiestaTM ProFusion III PSG monitor (Compumedics Telemed, Victoria, Australia) will be used to record the following parameters: electroencephalogram from three channels (F4/A1, C4/A1, O2/A1), bilateral electrooculogram, electromyogram of mentalis activity and bilateral anterior tibialis. Inductance plethysmography will be used to record respiratory movements of the chest and abdomen. Heart rate and electrocardiogram will be obtained from 2 chest leads. Oxygen saturation (SaO2) will be measured by an oximeter (Ohmeda Biox 3900 Pulse Oximeter) with finger probe. Respiratory airflow pressure signals and end tidal CO2 will be recorded simultaneously via a triple-port nasal catheter placed at the anterior nares and connected to a pressure transducer and a capnograph (BCI Capnocheck Plus) respectively. An oronasal thermal sensor will also be used to detect absence of airflow. Snoring will be measured by a snoring microphone placed near the throat. Body position is monitored via a body position sensor.

Respiratory events including obstructive apnoeas, mixed apnoeas, central apnoeas and hypopnoeas will be scored based on the recommendations from The AASM Manual for the Scoring of Sleep and Associated Events. [39] Respiratory effort-related arousals (RERAs) will be scored when there is a fall of <50% from baseline in the amplitude of nasal pressure signal with flattening of the nasal pressure waveform, accompanied by snoring, noisy breathing, or evidence of increased effort of breathing. A respiratory event is only scored when it lasts ≥2 breaths irrespective of its duration.

Obstructive apnoea hypopnoea index (OAHI), respiratory disturbance index (RDI), oxygen desaturation index (ODI), arousal index (ArI) and respiratory arousal index (RAI) are defined according to standard recommendations.

The investigators have carried out a pilot study to validate the proposed home sleep monitor (NOX-T3). Eleven subjects aged between 12-16 years were recruited from various sources, namely attendants to the sleep disorder clinic, healthy subjects who participated in our sleep duration and blood pressure study and relatives of patients who attend our respiratory clinic. Participants underwent an overnight in-laboratory polysomnography (PSG) with simultaneous NOX-T3 portable monitor recording, followed by unattended home recording with NOX-T3. The mean±SD OAHI was 5.8 ± 6.6 events/h on PSG, 6.0 ± 6.9 events/h on in-laboratory NOX-T3 recording, and 4.3 ± 4.8 events/h on home NOX-T3 recording (p=0.6). The intra-class correlation between OAHI obtained from in-laboratory PSG and in-laboratory NOX-T3 recording is 0.982 (95% CI: 0.937-0.995), indicating an excellent agreement. On the other hand, the intra-class correlation between OAHI obtained from in-laboratory PSG and home recording with NOX-T3 is 0.656 (95% CI: 0.155-0.893), indicating a moderate reliability of the home testing. Moreover, based on a cutoff of OAHI ≥1 or 5 events/h, the Cohen's Kappa agreement between in-laboratory PSG and home testing with NOX-T3 are both high at 0.814 (SE: 0.175).

The investigators plan to further validate this degree of agreement (Cohen's Kappa = 0.814) of the home sleep testing by recruitment more subjects during the study period to undergo both home testing and in-laboratory PSG. A sample size of 23 will provide an 80% power with a type I error of 0.05 to validate a Cohen's Kappa of 0.814 of the home sleep testing against the null hypothesis that the Cohen's Kappa is less than 0.3, i.e. a poor agreement. In other words, another 12 random subjects will be recruited during the study period, making it a total of 23 subjects for validating the home monitor.

7-day sleep pattern monitoring

Sleep will be objectively measured using actigraphs. Actigraph is a wristwatch-like device that uses a piezo-electric beam to detect movements which are translated into digital counts accumulated across predesigned epoch intervals (for example, 1 minute) and stored in internal memory. Data can then be downloaded for further analysis.

The device will be worn on the non-dominant wrist for 7 consecutive days and nights. The parents / caregivers will be instructed to keep a sleep log on bedtimes and waking times, temporary pauses in actigraph registration (for example, while taking a bath), and significant events that might affect sleep quantity or quality (for example, illness, injury etc). The child with help of parents / caregiver will be instructed to press a button (event marker) in the actigraph at bedtime and waking times. A completed sleep log will be obtained from all participants, including both parent-reported sleep log and event markers on bedtimes and waking times reported by the child. The activity data will be visually inspected to detect significant discrepancies among the sleep log, event markers, and the activity pattern. If there are discrepancies between event markers and sleep log recording, the latter will be used to define bedtime and waking times.

Recorded nights will be excluded if (a) the actigraph is not in use, (b) information on bedtime / waking times is missing, (c) there is a change in usual life routines due to, for example, illness.

The data will be scored using the Actiwatch Activity & Sleep Analysis software. The scored sleep data will be averaged for each study subject over the valid registration nights and separately for weekday and weekend nights. Sleep duration is defined as the actual sleeping time. Sleep start is defined as 10 minutes of consecutively recorded inactive data. Sleep efficiency is defined as the actual sleep time divided by the time in bed. Sleep duration and efficiency will be analysed as both continuous and dichotomous variables.

Upper Airway Size Examination

Tonsillar size will be examined and recorded using a validated grading system. Modified Mallampati score will also be obtained.

Conners' Continuous Performance Test (CPT)

CPT-II (Ver. 5.2 for Windows®) is designed as a computer game-like test for the evaluation of sustained attention and response inhibition for respondents aged 6 or above. Subjects are required to press the space bar or click the mouse whenever any letter except the letter 'X' appears on the computer screen. The test requires about 14 minutes to complete. It will be administered in the morning in a quiet room.

Determination of Levels of Biomarkers

Blood sample will be taken from the subjects for fasting (at least 8 hours) glucose level, lipid profile, high sensitivity C-reactive protein (hs-CRP), appetite-regulating hormones and pro-inflammatory cytokines determination. Plasma and serum samples will be stored at -80 °C until measurement. Plasma glucose is measured by hexokinase method (Cobas C8000 Clinical Chemistry Analyser, Roche Diagnostics Corp, Indianapolis, IN). The inter-assay coefficient of variation is 3% or less at all concentrations up to 41.6 mmol/L. The lipid profile includes total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) concentrations, which are measured by appropriate enzymatic assays (Cobas C8000 Clinical Chemistry Analyser, Roche Diagnostics Corp, Indianapolis, IN, USA) with inter-assay coefficients of variation of ≤3%. Serum hs-CRP will be measured by chemiluminescence immunoassay using the IMMULITE Analyzer (Diagnostic Products Corporation, Los Angeles, CA). Our reference range for hs-CRP is<11.0 mg/L and the sensitivity and inter-assay coefficients of variation are 0.1 mg/L, 6.6% at 1.7 mg/L, 6.5% at 4.8 mg/L and 5.7% at 7.1 mg/L. Ghrelin, leptin, adiponectin and proinflammatory cytokines TNF-α and IL-6 will also be measured using Enzyme-Linked Immunosorbent Assay.

Inclusion Criteria:

1. Adolescents aged 12-16 years
2. Written informed consent from parents and assent from participants available

Exclusion Criteria:

1. Previously undergone upper airway surgery or currently receiving treatment for OSA
2. Craniofacial anomalies
3. Unable to speak or read Chinese

Adolescents will be recruited from local secondary schools that are randomly selected from the 4 geographic regions of Hong Kong. The sample selection will be based on a stratified (by districts) and clustered (subjects within randomly selected schools) randomised sampling frame. We stratify all secondary schools registered under the Education Bureau according to the four geographic regions in Hong Kong (i.e. Hong Kong Island, Kowloon, New Territories East and New Territories West). In order to avoid sampling bias towards the region that has less population, the final number of schools to be recruited from each of the four geographical regions would be in proportion to the adolescent population that reside in that region.